Randomized Experiments for Online Content Moderation Policy

The difficulty of achieving consensuses over the regulation of online content moderation has created a stringent and divergent regulatory framework around the world. This fragmentation increases the cost of operating in digital global markets due to greater entry and expansion barriers. Given that countries’ legal standards over the regulation of content moderation remain too far apart from each other, international law does not seem to offer a solution in this respect. Nonetheless, policymakers and researchers could start taking advantage of the divergent legal environment and the data richness that characterizes the digital economy to test and prove the effects of the different regulations in place. Today, opinions and proposals are based to a great extent on intuitive assumptions or theoretical ideas. Thus, it is necessary to start questioning and testing those ideas through empirical methods. Applying the rationale used in randomized experiments to validate the intuitive assumptions that fill the debate, such as the alleged effects of intermediary liability, including the chilling-effect over speech, would allow policymakers to better understand how the different regimes shape the conduct of intermediaries and make policy decisions accordingly.

By Imanol Ramírez^[1]

I. INTRODUCTION

Much ink has been spilled on how to balance the different policy objectives of online content moderation regulation. These policy goals include preventing online harms, promotin